Research Note vs Research Report: What Each Carries

A research note and a research report differ in what they carry, not in how good they are. A note answers one question about one event and assumes the reader already holds the view it sits inside. A report carries the whole view: the beliefs it rests on, the evidence, the peer set with its exclusions, the risks, and what would change it. A note read as a report is read without its assumptions.

Three things sit underneath that and none of them is rebuilt here. The contents of each shape are set out separately. Sorting the two against each other is a different job from describing either one again. From the thesis material comes the idea of an assumption setThe short numbered list of beliefs a written view depends on, each one stated plainly enough that somebody could go and disagree with it. Built in the thesis sequence., a short numbered list of beliefs a view depends on, and it is used here rather than constructed. And from the shared teaching record comes Sarvani Coatings Limited, an invented listed maker of decorative paints and industrial coatings, whose published year three figures give the worked instance below something concrete to be about.

What is a research note, taken on its own terms?

A comparison made before both sides are defined is just a list of adjectives, so start with each shape by itself. A note is a short published piece that answers one question about one thing that happened. A result came out. A plant caught fire. A rival cut prices. The company said something at an investor meeting. The note takes that one occurrence, states what it was, states what the analyst thinks it means for a view already on record, and stops.

The defining property is not the length. A note is written on the assumption that the reader already holds the view it sits inside, so the note supplies only the new piece and none of the standing structure. Everything the reader needs in order to place the new piece has already been published somewhere else, and the note quietly relies on that.

Consider a message from a building society secretary saying the lift is back in service from Tuesday. The message is complete and useful and takes eleven words. Everybody reading it already knows there is a lift, that it was broken, and roughly what the repair was going to involve. The same eleven words sent to somebody who has never been in the building teach nothing at all. The message was never self contained. The notice was an increment to a shared situation, and its shortness follows from that rather than standing as an achievement.

What is a research report, taken on its own terms?

A report is a published piece that states a whole view and everything the view depends on. A report says what the analyst thinks, why, on what evidence, against which comparable companies and with which ones deliberately left out, what could go wrong, and at what point the analyst would stop believing it. An initiationThe first published piece an analyst writes on a company they have just begun following, which is why it has to state everything from the beginning rather than assuming a reader. is the clearest instance. By definition it has no prior reader, so nothing at all can be assumed.

A report is written to survive a reader who has never seen anything else the writer has produced, and that single requirement is what forces everything else about its shape. If a stranger has to be able to follow it, the beliefs have to be written down. If the beliefs are written down, they can be argued with. If they can be argued with, the writer has taken on something real.

Notice what has just been established and what has not. Both shapes are the analyst's own published work. Neither is a draft. Neither is more careful than the other. The two shapes differ in how much of the reasoning has to be inside the covers, and that difference is one of scope of commitment, not of effort or quality.

What sets each one going?

A note is triggered by something that happened outside the analyst. The event arrives on its own schedule, the analyst did not choose it, and once it has happened the coverageThe standing undertaking an analyst takes on to keep following a named company and to keep publishing on it, rather than writing about it once. Settled at the start of this sequence. commitment settled at the start of this sequence means a piece has to go out. Results come every quarter whether anyone feels ready or not.

A report is triggered by a decision the analyst made. Two decisions, mostly: to begin following a company that was not being followed, or to rebuild a view that has drifted so far from the document describing it that the document is no longer a fair statement of what the analyst thinks. Both are voluntary.

Nothing in the world forces a report to be written. A view can run for years without one, and nobody notices until somebody asks what the assumptions were. That asymmetry is the whole of this axis. Events push notes out at a steady rate. Nothing pushes a report out at all.

The household version is a standing order. Money leaves the account every month because the date arrives, and it will keep leaving on that schedule for as long as nobody intervenes. Sitting down to work out whether the thing being paid for is still wanted is a decision nobody schedules. The payments are the notes. The sitting down is the report, and it is the one that never happens by itself.

What does each assume the reader already holds?

A note assumes the reader holds the view and needs only the increment. A report assumes nothing and has to build from the ground. The distinction sounds obvious written down. A document does not announce which of the two it is being, so the distinction is almost never applied.

So here is a test that costs nothing. Hand the document to somebody competent who does not follow the company, and see whether they can state what the writer believes and why, using only what is in their hand. If they can, it is a report. If they come back with a conclusion and no idea what it rests on, it was a note, whatever it was labelled and however long it ran.

The test is worth running because the two natural proxies both fail. Length fails: a note can run to eleven printed sides of careful reasoning about one supplier contract, and a report on a simple business can close in nine. The label fails too. The word on the cover is chosen by whoever formats the document, and that person is often not the analyst.

What does each commit the writer to?

A note commits the writer to one claim about one event, and to nothing else. If the claim is right the note was right, and no other sentence in it is on the line. A report commits the writer to every number in it, including all the ones nobody will ever ask about. The forecast three years out that no reader reaches. The line in the peer table explaining why a company that looks comparable was left out. The sentence naming the level at which the writer would give up.

Committing to all of them is why a report is uncomfortable to write, and why writing one is the honest act: it puts every assumption somewhere a reader can go and contradict it. The discomfort is the mechanism working, not a sign that something has gone wrong.

The same thing shows up outside finance. Saying a scooter is a good buy commits the speaker to very little. Writing down that it is expected to do fifty kilometres to the litre, need one service a year and hold half its value after three, and that the judgement would count as wrong if it did any worse, commits the speaker to three checkable things and a stopping rule. Nobody can prove the first statement wrong. Anybody can prove the second wrong, and that is exactly what makes the second worth reading.

How long is each one good for?

A note expires with its occasion. The result it discussed stops being the latest result, the price move it explained is three months old, and the note quietly becomes a dated record of what somebody thought at the time. The occasion is printed on the front and the reader can see the calendar, so nobody is confused by a note going stale.

A report expires when one of its assumptions moves. Expiry might come a week later or two years later. Whatever expired a report sits outside the document, so the report itself cannot establish whether expiry has happened. A four month old report can be perfectly current and a four week old one can be dead. The date on the cover carries almost no information about either.

The fix is the run of dated notes published behind a report, each one saying which assumption the new event touched and whether it still holds. The run does something a date can never do: it is evidence that somebody has looked recently. A report standing alone with nothing published behind it is a historical artefact wearing the clothes of a current view.

What can a reader safely take from each?

The reasoning and the assumptions are physically inside a report, so take both. From a note, take the fact it reports, and take nothing about the view. Taking nothing about the view is where careful readers still go wrong, and the mechanism is small enough to miss.

A note that says a result confirms the view has used a verb whose meaning lives in a document the reader may never have opened. The word does real work. The verb asserts that a specific stated belief has been tested against a specific new observation and has survived. Every part of that assertion sits in the other document, and a reader without the other document has no way of telling whether the verb was earned or simply typed.

What can a sentence be, if every word in it is true and it is still unusable?

Here is one sentence, lifted out of a note on Sarvani Coatings Limited written by Meghna Iyer, the analyst whose work this sequence follows. The sentence reports one quarter's EBITDAThe letters stand for earnings taken before interest, before tax, and before depreciation and amortisation. A profit struck partway down the ladder, after materials, staff and running costs, with the financing and asset write-down lines still to come. Settled in the accounting layer., meaning earnings before interest, tax, depreciation and amortisation, and sets it against the year. Every figure in it comes from the invented teaching record, and every one of them is checkable.

Check it. Rs 107 crore over Rs 580 crore is 18.4483 per cent and prints as 18.45. The year's Rs 446 crore over Rs 2,415 crore is 18.4679 per cent and prints as 18.47. The four quarters sum to the published year on both lines. Nothing in the sentence is wrong, and one small thing in it is already doing quiet work: differencing the two unrounded results gives 0.0196 of a point, so the quarter came in fractionally below the blended year rather than level with it. The word "against" carries that without saying it.

The real problem has nothing to do with rounding. The assumption the sentence claims to confirm is about the gross marginWhat is left of a rupee of sales once the input cost of the goods sold has been taken out, before any other expense is counted at all. Settled in the accounting layer.. The figure the sentence reports is an EBITDA margin. The gross margin and the margin the sentence reports are two different lines, with two whole cost blocks sitting between them.

Rs 259 crore is 10.72 points of revenue on its own. Advertising can be pulled forward into a festive quarter or pushed back into the next year. Freight moves with where the goods went and when. Either can swing a quarter's EBITDA margin by more than the whole difference the note is calling steady, without one paisa of movement in the gross margin line.

The mismatch gets worse, and the worse half is the part worth carrying away. Suppose a reader decides to go and check the gross margin for the quarter directly. The check cannot be done. The release carries revenue and it carries EBITDA. The release never breaks out the materials line by quarter, so no gross margin for the fourth quarter can be read from the record or built out of it. The note's verb cannot be tested from the release it was written about, not by a careless reader and not by a careful one.

Look instead at how far a quarterly EBITDA margin wanders on its own. The spread settles how much weight a single quarter can carry.

Open the report the note points at, and the assumption stops being a word. The assumption reads: gross margin holds while realisation per unitThe average price a maker actually got for each unit it sold, found by dividing revenue by volume rather than by reading a price list. grows faster than materials cost per unit. A sentence in that form can be tested. Over the one year from year two to year three, on volume up 6.0 per cent, realisation per unit rose 7.47 per cent while materials cost per unit rose 3.64 per cent, so realisation outran cost by 3.83 points of growth and the gross margin widened by 1.99 points, from 44.01 to 46.00. The record prints those as about 7.5, about 3.6 and 2.0.

Two disciplines are being held there at once. The per unit pair is a one year decomposition and belongs beside the one year gain of 2.0 points; it must never be attached to the two year headline. And the arithmetic is forced rather than confirmed: volume cancels exactly out of the ratio, so the materials share has to move by the ratio of the two per unit growth rates. The agreement is algebra closing, not a second piece of evidence.

The assumption also has a flip pointThe level at which a stated assumption stops supporting a conclusion and starts contradicting it. Worked out in detail in the post mortem material that follows this sequence.: realisation growth of 3.64 per cent, the rate at which it exactly matches materials cost per unit and the gross margin comes out flat. Above it the margin widens, below it the margin narrows. Now, at last, the note's claim is a claim that can be tested. And it fails. The note offered an EBITDA margin for a single quarter against an assumption stated in per unit terms over a year, on a line the quarterly release does not even carry.

A note's verbs are only as good as the document they point at, and a reader who never opens that document is trusting a word rather than reading a claim. The note reported a true figure, attached a true label and confirmed nothing, and only the report makes that visible.

When is a note the right shape, and when is it avoidance?

A note is right when two things are true at the same time. The reader genuinely holds the view, and the increment is genuinely small. Both halves matter, and they fail in different ways. If the reader does not hold the view, a note is unreadable however small the increment. If the increment is not small, a note is a compression of something that needed room.

A note becomes avoidance when the analyst uses the short shape to state a conclusion whose reasoning they have never written down anywhere. Avoidance of that kind is not laziness, and from the inside it rarely feels like dishonesty. Writing nothing down feels like efficiency. The view is in the analyst's head, it has been there a long time, it has been discussed in meetings, and putting it on paper feels like paperwork rather than work.

The honest test is not about the note at all: it is whether the report the note points at actually exists and is current. If it does, the note is doing its job. If it does not, the note is a conclusion with no reachable basis, and no amount of care in the writing of the note can supply one.

What do the two shapes have in common?

A reader could easily come away thinking a report is serious work and a note is not. The differences sorted above are all about scope and none of them is about care, so that conclusion is the wrong one.

Both carry a date. Both carry the declarations the writer is required to make about their own position and interest. Both go out under the writer's own name, into the same public record, with the same consequences if a figure in them is wrong. A wrong number in a note of two printed sides is exactly as wrong as a wrong number in a report of forty, and the writer is exactly as answerable for it.

The comparison sorts these two documents; it does not rank them, and the useful question is never which is better but which one the claim being made actually requires. A claim about one event needs a note. A claim about a company needs a report. Reaching for the wrong one is a mistake whichever direction it goes in, and a report of forty printed sides about a single supplier contract is its own kind of failure.

The two shapes differ by what triggers them, what they oblige the writer to, and how much of the reasoning has to be inside the covers, and none of that lies along a continuous quantity. A spectrum between a note and a report is not something the practice supports. The one quantity anywhere near this subject that genuinely moves, the level at which the margin assumption reverses, belongs to the post mortem material further along and is built there.